In the past year, most of the major independent BI suppliers have been gobbled up. Things got going in February with Oracle’s acquisition of Hyperion. Then in October SAP purchased Business Objects. Finally, IBM entered the foray when it bought Cognos in November. Aside from the specialist statistical software suppliers like SPSS and SAS, there is hardly a standalone BI vendor left.
There have probably been two major catalysts for all this frenzy. The first of these is BI’s abysmal track record. Almost the first promise IT gave the business was information at your fingertips and for many businesses their executive are still waiting. The second of these are the moves by Microsoft to embed more advanced BI functionality within Office applications.
BI has masqueraded as a succession of names in the course of its history. Its first manifestations were as a query language which sat on top of an application and enabled a user to interrogate the data being captured. In the early 1990’s these evolved in to executive information systems which encapsulated more sophisticated query and reporting tools. However, soon businesses began to realize that interrogating application data in real time was a massive drain on processing resources which seriously impacted the performance of core business systems. As such, data warehousing applications emerged where data was transferred to separate databases where it could be catalogued and interrogated offline. The BI applications of today combine data warehouse capabilities with advanced analysis functionality.
As these systems have evolved businesses have gained greater insight in to the challenges of utilizing BI. Above all they have found the biggest issue is data quality. Often separate business systems serve different business units and the meaning of common data terms within these systems can vary widely. One senior IT executive at a major Australian bank once advised me that at least 70 percent of the effort and expense incurred in establishing a data warehouse was in the task of cleansing the data to iron out these disparate definitions between systems. The Data Warehouse Institute (TDWI) reports that 83 percent of respondents to a recent survey reported their organization had suffered problems because of poor data management. The dominant problems were inaccurate reporting and arguments over which data to trust.
ERP to BI
ERP sought to eliminate data fragmentation by creating a comprehensive suite of integrated business systems. However, it has proved impossible to provide an ERP set that addresses all facets of an organization’s operations. Therefore, the move by the major ERP vendors into the BI space probably reflects a desire by these suppliers to solve the problem of disunity in corporate data. The new approach is to use BI tools to achieve results at the back end. This approach leverages the 80/20 rule since much of the data definition work could be a standard element of an ERP implementation or upgrade.
However, the next challenge becomes training the users to harness the BI applications. In many cases users can find IT training less than satisfactory as it is theoretical rather than structured around examples taken from their actual work. As such, many organizations tend to under invest in IT training hoping their users will learn on the job after a new system has been installed.
However, there are applications where users are proficient. Typically, these are the components of the Microsoft Office suite. Users are usually very open to further training in these applications because they tend to be common across businesses. As a result, a user knows that a good understanding of these programs is important as the skills gained will be transferable to any new job they take. This attitude is particularly prevalent towards Excel which has become the de facto analysis tool for most senior business executives.
Excel over-rules
Many BI suppliers will say that their biggest challenge is getting users away from Excel as they want to use this to interrogate corporate data. In fact it is not uncommon for users to download data from a data warehouse to a spreadsheet for analysis and manipulation. Unfortunately, this defeats an investment in BI. It also means that important business decisions are being made on obsolete data since spreadsheets are usually not as frequently updated as data warehouses. This lack of currency of data in spreadsheets is behind the observation that 90 percent of the data in spreadsheets is wrong. This became such a issue for a major retailer in New Zealand that they ended up stamping ‘certified true’ across all reports generated by the data warehouse. In any corporate dispute over the facts a report produced from the data warehouse always took preference over information provided from a spreadsheet.
However, Microsoft is now making a commitment towards a new range of BI offerings. In addition, it has also acquired the Officewriter product, a tool that will provide BI report writing capability within Excel. The promise is that users will have BI functionality available to them within Excel which negates the need for them to learn a new product. Moreover, they can get this functionality without having to leave a product where they spend a good proportion of their working day. The result could be that BI becomes as mainstream as other
Office applications like Word and PowerPoint. If so then the ERP vendors could find that they lose ownership of the corporate data to Microsoft. If this were the case then the ERP vendors’ prominence within the corporate IT department could diminish. Hence the ERP vendors move in to the BI space could indicate a defensive strategy against Microsoft’s BI intentions.
Microsoft does not have the best track record of delivering on time. As such, the ERP suppliers probably have sufficient time to integrate and consolidate their BI acquisitions before Microsoft becomes a force in the BI arena. Nevertheless, CIOs should not be troubled by any of these corporate machinations. It is early days and the real results of all these BI acquisitions and product announcements will take time to materialize. Therefore, any BI requirements they have should be assessed against what is the best serves their current needs, rather than what might be available in the future once the BI market has settled down.
Peter Hind is a consultant with years of experience in the IT industry.
